From Bejeweled to Triple Tile: Why Match-3 Never Gets Old

Match three of a kind. That is the entire rule, and it has been the entire rule for about thirty years — through gems, candies, farm animals, souvenirs, blocks with faces and, in our case, cars. Genres normally exhaust themselves; this one keeps getting rebuilt from the studs and sold again to the same delighted people. Here is the lineage, honestly told, and the reason the number three does so much of the work.

The mid-90s: it starts as Russian shareware

The commonly cited origin is Shariki, a DOS game written by Eugene Alemzhin in 1994 and passed around as shareware. Swap two adjacent balls, line up three of a colour, watch them vanish and everything above fall into the gap. Every load-bearing piece of the genre is already sitting there in that first game: the swap, the clear, the collapse, and the cascade when the falling pieces happen to line up again on their own.

It made nobody rich, which is a fairly standard opening chapter for a genre. Inventing a mechanic and owning a mechanic turn out to be very different jobs.

2001: Bejeweled hands it to everyone

PopCap released Bejeweled in 2001 — first as a game you could play in a web browser, later as the downloadable version most people remember. The rules were Shariki’s. What was new was everything around them: the gems glinted, the chime rose in pitch as your cascade ran, and a voice told you that you were excellent. It ran on absolutely anything with a screen, which in the years that followed included a great many early mobile phones.

This is the moment match-3 stops being a mechanic and becomes a format — something other developers build in, rather than a specific game people play.

Shariki invented the rule. Bejeweled invented the feeling — and it turns out the feeling is the part you can sell.

2007: Puzzle Quest bolts an RPG onto the board

For years the ceiling on a match-3 was the high score. You played well, the number went up, you closed the window. Puzzle Quest: Challenge of the Warlords, released in 2007, changed the question from “how well did you play?” to “what did that get you?” Gems became mana, skulls became damage, and the board sat inside a campaign with a character, spells and enemies who had their own tricks. You were still matching three. You were now also levelling up.

It is arguably the single most consequential idea in the genre’s history: the matching does not have to be the game — it can be the engine for a game that is about something else entirely. Almost everything that followed is a variation on that sentence.

2012: Candy Crush makes the level the unit

Candy Crush Saga, from King in 2012, took the Puzzle Quest insight and pointed it at a map. Instead of a campaign you got a numbered path of hand-authored levels, each with its own objective — clear the jelly, drop the ingredients, hit a score in twenty moves. Lives refilled over time. Boosters were available for money. Your friends were visible on the same winding road, a few levels ahead, being smug.

Whatever you think of that economy, it is worth being precise about what actually changed. Before Candy Crush, a match-3 was mostly a machine you fed until you lost; the session was the unit. After it, the level was the unit — an authored puzzle with intent, a goal, a move limit, a failure state and a place on a difficulty curve. That is a completely different kind of design work, and it is why the genre suddenly needed hundreds of designers instead of one very good board.

The 2020s: the tray arrives

The newest branch drops the swap altogether. You tap a piece off a crowded, often 3D board; it goes into a small tray at the bottom of the screen; three of a kind in the tray clears out. Fill the tray with mismatches and the run is over. These are the triple-tile or triple-match games, and they broke out in the early 2020s.

Their family tree runs less through Bejeweled than through mahjong solitaire — the layered pile where only the free pieces on the edges can be taken. What the tray adds is pressure of a very modern kind: the board’s congestion becomes the puzzle, and your own greed becomes the threat. Nothing on the board can hurt you. Only what you have already picked up can.

Why three?

Not two, not four, not five. There are two good reasons, and neither is arbitrary.

The first is perceptual. People can take in very small quantities — up to about four items — instantly, without counting them one by one; psychologists call it subitizing, and it is one of the more solidly established facts about how we read a scene. Three sits comfortably inside that window. Five does not. A match-3 board can therefore be read at a glance rather than parsed, which means the whole game runs at the speed of looking, and that is why nobody has ever needed a tutorial for one.

The second is economic, in the design sense. Every matching game makes an implicit promise: there will always be something to do. The bigger the group you require, the more expensive that promise gets — a board that must always contain a findable set of five is a much more constrained board than one that needs a set of three. Meanwhile a match-2 is no promise at all, because any pair is a match and nothing is a puzzle. Two is a coincidence; three is a set. Three is the smallest number that reads as a deliberate pattern and the largest number a game can afford to guarantee almost everywhere at once.

Near misses and free money

Two things keep the loop turning, and both are plainly observable in your own head while you play. The first is the near miss: two of a kind sitting in the tray, or a board one square away from a chain. A near miss is a much stronger pull than a blank board, because you know exactly what would have worked and how small the gap was. “Almost” is one of the most engaging states a game can put you in.

The second is the cascade — the clear you planned dropping pieces you did not plan into a match you did not earn. The board pays you for something you did not do, unpredictably. Rewards that arrive on an irregular schedule hold attention better than rewards that arrive on a reliable one; that is well established, it is not an accident, and it is why cascades survived intact into every descendant of Shariki. Worth knowing the choice is being made, even when you are enjoying it.

Why it keeps reinventing itself instead of dying

Most genres die of exhaustion: the mechanic gets fully explored and there is nothing left to say. Match-3 escapes because the rule is not the game — the rule is the verb. Matching three is a container, and it will hold whatever you pour into it: an RPG campaign, a saga map, a house to renovate, a farm to rebuild, a garage of cars to restore, a heist to pull. The matching is the metronome. Everything interesting happens in the layer built on top of it, which is why the genre can be reinvented indefinitely without anyone touching the core loop.

It is also why arguments about match-3 games are almost never about matching. Nobody complains that three gems is too few. They talk about lives, timers, difficulty spikes and prices — all of which live in the meta layer, not the mechanic. The mechanic was finished in 1994.

The family tree, briefly

Play the modern branch

We build in the newest part of this tree, so naturally we would like you to try it. Both of these are free in the browser, with no download and no sign-up:

• Travel Tiles — triple-tile matching in its purest form: tap souvenirs into a six-slot tray and clear all ninety-six, untimed, at your own pace.
• Car Match — the same tap-to-match mechanic laid over a traffic grid, where only the reachable cars can be picked and clearing a path is half the job.

Thirty years on, nobody has improved the rule, because there is nothing left to improve. Three of a kind, and they go away. Everything since has been costume, context and a better chime — which is, it turns out, exactly enough to keep a genre alive.